Discussion Closed This discussion was created more than 6 months ago and has been closed. To start a new discussion with a link back to this one, click here.
Convergence Problems
Posted Feb 12, 2020, 8:27 a.m. EST Structural & Acoustics, Results & Visualization, Structural Mechanics 1 Reply
Please login with a confirmed email address before reporting spam
Hi, everyone! I am trying to simulate the contact between tire and soil in COMSOL, but I am getting too many problems of convergence. The errors are always like:
Failed to find a solution for the initial parameter. Maximum number of dogleg iterations reached. There was an error message from the linear solver. The relative error (0.18) is greater than the relative tolerance. Returned solution is not converged. Not all parameter steps returned.
The geometry is relatively simple, because it was simplified to 2D. However, some of the results that we could get an answer without the convergence error, don't represent what occurs in real life or at least a reasonable result. I am afraid that the boundary conditions may not be set properly. I am new to the software and still getting the hang of it. My modelation is as follows:
Initially, the geometry is basically a circle (tire/wheel) on top of a rectangle (soil). I used a contact pair between the tire and soil, used Solid Mechanics as my physics module, fixed constraint for the soil, added mass, gravity, assumed a hyperelastic Material, Soil Plasticity for the soil and also Spring Foundation. In the contact pair, I avoided adding friction to simplify the problem and also avoided to add Prescribed Displacement so that the displacement would be generated by the weight of the tire to the soil. Any idea why these errors are occurring? Thank you for your attention Matheus S.
Please login with a confirmed email address before reporting spam
I only have one comment on the convergence. The built-in convergence criterion is 1e-3, right? I doubt how much sense this value makes, considering the real physical situation can be very different. So my experience is that one DO NOT need to care about it. Just ignore the convergence error and proceed. As for if your results are truely converged or not, one can always do convergence study oneself by redoing the simulation with a series of parameters. One should not rely too much on such a general purpose convergence judgement.
-------------------ZHANG, Pu
School of Physics,
Huazhong University of Science and Technology